arXiv subjects

Ali Bereyhi

Publications and source records attributed to Ali Bereyhi.

At least 19 recordsLinked to original sources

arXiv2026

Evolution of Financial AI: From Statistical Models to Agentic AI

Artificial intelligence (AI) in finance has evolved from statistical learning and machine learning for structured prediction, through deep learning and foundation models for representation learning, language understanding, and reasoning, to retrieval-augmented generation for knowledge-grounded reasoning and agentic AI for autonomous planning and execution. In this paper, we examine how researchers and practitioners can select and apply AI methods in finance. We first show that the major paradigms of financial AI address fundamentally different classes of problems, rather than forming a strict hierarchy of capability. We then review the current state of research on AI in finance and trace how these paradigms have reshaped financial data, financial tasks, decision support, and human-AI collaboration. Building on this, we propose practical guidelines for matching AI paradigms to task characteristics, and show that more advanced models, such as foundation models or autonomous agents, are not universally superior. We conclude by outlining open challenges for researchers and practitioners working toward reliable financial AI.

cs.CE↗

arXiv2026

Understanding Large Language Models as Numerical Optimizers

Large language models (LLMs) have emerged as potential tools for numerical optimization. Despite promising results, their intrinsic optimization capabilities and limitations remain poorly understood, particularly in high dimensions. This paper investigates the ability of LLMs to exploit mathematical problem structures and generate feasible, near-optimal solutions. We develop an evaluation framework that distinguishes solution quality from execution failures, including numerical inaccuracies, constraint violations, and output-format errors. Using representative problems from convex optimization, fractional programming, and constrained projection, we assess LLM performance against deterministic reference solvers across different dimensions. Our results reveal that LLMs can capture useful optimization structures and produce near-optimal solutions, even when their outputs fail to satisfy numerical or formatting requirements. However, reliability deteriorates with increasing problem dimension and depends strongly on the underlying optimization structure. To further investigate the limitations, we deploy an iterative refinement mechanism that provides targeted feedback on feasibility and solution quality. The results demonstrate that iterative refinement substantially improves reliability and extends the range of problem dimensions over which LLMs produce valid solutions. Nevertheless, persistent numerical and structural errors limit their applicability as standalone optimizers. These findings provide insights into the capacity and limits of LLM-based optimization and motivate hybrid approaches that combine LLM reasoning with deterministic numerical verification.

cs.AI↗

arXiv2026

Revisiting GAN with Bayes-Optimal Discrimination

We approach generative adversarial network (GAN) training from its original real-versus-fake classification viewpoint, replacing the cross-entropy discriminator objective with an upper bound on the Bayes error rate (BER). Using the Bayes optimal learning threshold (BOLT) loss, we formulate an adversarial game in which the discriminator tightens this bound and the generator maximizes it. We characterize the unrestricted optimal discriminator and show that, under balanced class priors, the population game is equivalent to minimizing total variation between the data and generator distributions. Restricting the bounded discriminator to be 1-Lipschitz yields BOLT-GAN. For balanced priors and the full discriminator class, we establish upper and lower bounds relating its population discrepancy to Wasserstein-1 distance on bounded domains. Controlled experiments on four image-generation benchmarks show improved sample quality and coverage over WGAN-GP under matched architectures, optimization settings, and gradient penalties. At a common 20-epoch budget, BOLT-GAN reduces Fréchet Inception Distance (FID) from 60.0 to 44.2 on CIFAR-10 and from 102.5 to 28.4 on LSUN Bedroom-64.

cs.LG↗

arXiv2026

BOLT: Universal Classifier Training via a Bayes-Error Upper Bound

This work introduces a novel approach to compute a universal upper bound on the Bayes error rate (BER) from classifier-output samples, which serves as a Bayes-error-related performance criterion for the underlying classification task. We show that the bound is tight in the binary case and characterize the max-sum gap that appears in the aggregate multiclass formulation. Based on this bound, we define a new loss function, called Bayes optimal learning threshold (BOLT), for training parameterized classifiers. For each data sample, the BOLT loss computes a term whose expectation upper-bounds the Bayes error, so that during training the model parameters can be updated toward minimizing this bound. Numerical experiments with both synthetic and real-world datasets show that, in the reported settings, BOLT training often attains lower held-out classification error than cross-entropy training.

cs.LG↗

arXiv2026

Fast Tuning of Pinching Antenna-Assisted ISAC Systems via User Clustering

Pinching-antenna systems (PASSs) offer a great degree of flexibility for integrated sensing and communications (ISAC) by adapting the pinching locations for both communication channel reshaping and sensing-beam steering. Despite that, direct location optimization in PASSs is high-dimensional and non-convex, making their fast reconfiguration difficult in practice. This work proposes a low-complexity PASS-assisted ISAC design based on user clustering. Instead of direct optimization of pinching locations, the elements on each waveguide are distributed over multiple segments, where each segment serves a cluster of users or targets. This alternative configuration reduces the location tuning problem to a low-dimensional weight allocation task, which can be efficiently addressed via a gradient based method. We propose a low-complexity iterative algorithm, in which the digital precoder is updated via an equivalent weighted mean squared error minimization, and pinching locations are updated via weight allocation and projection. Our numerical experiments show that the proposed design can significantly enhance the communication-sensing trade-off of ISAC systems over the benchmark designs with modest computational cost. The results highlight the potential gain of deploying PASS-assisted transceiver in practical ISAC systems.

eess.SP↗

arXiv2026

Global Unknown Estimation: A Statistical Framework for Wireless Distributed Learning

Over-the-air computation (AirComp) is widely used for model aggregation in wireless distributed learning. Although it enhances communication efficiency, we believe the AirComp aggregation has limited effectiveness due to the difference between its target problem and that of distributed learning. In this paper, we develop a rigorous formulation for optimal model aggregation in wireless distributed learning. Using this formulation, we show that AirComp aggregation generally assumes a mismatched statistical model for local parameters. We then propose a statistical framework for model aggregation, called global unknown estimation (GUE). It captures the statistical relation between the local and global model parameters, allowing to interpret model aggregation as an inference task. We validate the efficiency of GUE through numerical experiments. Our results show that, in the low SNR regime, GUE can reduce the required power for model aggregation by approximately 15 dB compared to AirComp aggregation. Remarkably, this gain is obtained without additional computational overhead

eess.SP↗

arXiv2026

LDGuid: A Framework for Robust Change Detection via Latent Difference Guidance

Modern deep learning models for change detection (CD) often struggle to explicitly represent task-relevant semantic differences. This paper proposes the Latent Difference Guidance (LDGuid) framework that explicitly learns and injects semantic differences into CD models. LDGuid deploys adversarial autoencoding to implement a difference embedding (DE) module. The DE module is pretrained via the information bottleneck method, restricting it to learn only task-relevant differences between pre- and post-event samples. The learned latent difference is then used as an explicit guidance signal in the CD model. We validate LDGuid by integrating it into U-Net, BIT, and AERNet baselines for CD and evaluating it on LEVIR-CD, WHU-CD, SVCD, and CaBuAr datasets. Experimental results show that LDGuid enhances segmentation performance across all benchmarks, with particularly remarkable gains in challenging settings affected by spectral noise. The results further highlight the ability of LDGuid in incorporating domain knowledge, such as task-specific spectral indices. Our findings suggest that semantic difference learning can drastically enhance the robustness of CD in remote sensing.

cs.CV↗

arXiv2026

Low-complexity tuning of pinching-antenna systems for integrated sensing and communication

Pinching antenna systems (PASSs) can dynamically adapt their transmit and receive arrays for sensing and communication in wireless systems. This work explores the potential of PASSs for integrated sensing and communication (ISAC) by proposing a novel PASS-aided ISAC design, in which pinching locations are adaptively adjusted to enable simultaneous sensing and data transmission with minimal interference. The proposed design introduces a bi-partitioning strategy that allocates sensing power and tunes pinching locations with remarkably low computational complexity, allowing dynamic PASS tuning at high update rates. Numerical results demonstrate that the proposed approach achieves a significantly larger sensing-communication rate region compared to baseline designs at no noticeable cost.

cs.IT↗

arXiv2026

Regime-aware financial volatility forecasting via in-context learning

This work introduces a regime-aware in-context learning framework that leverages large language models (LLMs) for financial volatility forecasting under nonstationary market conditions. The proposed approach deploys pretrained LLMs to reason over historical volatility patterns and adjust their predictions without parameter fine-tuning. We develop an oracle-guided refinement procedure that constructs regime-aware demonstrations from training data. An LLM is then deployed as an in-context learner that predicts the next-step volatility from the input sequence using demonstrations sampled conditional to the estimated market label. This conditional sampling strategy enables the LLM to adapt its predictions to regime-dependent volatility dynamics through contextual reasoning alone. Experiments with multiple financial datasets show that the proposed regime-aware in-context learning framework outperforms both classical volatility forecasting approaches and direct one-shot learning, especially during high-volatility periods.

cs.LG↗

arXiv2026

IRIS: Learning-Driven Task-Specific Cinema Robot Arm for Visuomotor Motion Control

Robotic camera systems enable dynamic, repeatable motion beyond human capabilities, yet their adoption remains limited by the high cost and operational complexity of industrial-grade platforms. We present the Intelligent Robotic Imaging System (IRIS), a task-specific 6-DOF manipulator designed for autonomous, learning-driven cinematic motion control. IRIS integrates a lightweight, fully 3D-printed hardware design with a goal-conditioned visuomotor imitation learning framework based on Action Chunking with Transformers (ACT). The system learns object-aware and perceptually smooth camera trajectories directly from human demonstrations, eliminating the need for explicit geometric programming. The complete platform costs under $1,000 USD, supports a 1.5 kg payload, and achieves approximately 1 mm repeatability. Real-world experiments demonstrate accurate trajectory tracking, reliable autonomous execution, and generalization across diverse cinematic motions.

cs.RO↗

arXiv2026

Dynamic and Static Energy Efficient Design of Pinching Antenna Systems

We study the energy efficiency of pinching-antenna systems (PASSs) by developing a consistent formulation for power distribution in these systems. The per-antenna power distribution in PASSs is not controlled explicitly by a power allocation policy, but rather implicitly through tuning of pinching couplings and locations. Both these factors are tunable: (i) pinching locations are tuned using movable elements, and (ii) couplings can be tuned by varying the effective coupling length of the pinching elements. While the former is feasible to be addressed dynamically in settings with low user mobility, the latter cannot be addressed at a high rate. We thus develop a class of hybrid dynamic-static algorithms, which maximize the energy efficiency by updating the system parameters at different rates. Our experimental results depict that dynamic tuning of pinching locations can significantly boost energy efficiency of PASSs.

eess.SP↗

arXiv2026

Energy-Efficient Over-the-Air Federated Learning via Pinching Antenna Systems

Pinching antennas systems (PASSs) have recently been proposed as a novel flexible-antenna technology. These systems are implemented by attaching low-cost pinching elements to dielectric waveguides. As the direct link is bypassed through waveguides, PASSs can effectively compensate large-scale effects of the wireless channel. This work explores the potential gains of employing PASSs for over-the-air federated learning (OTA-FL). For a PASS-assisted server, we develop a low-complexity algorithmic approach, which jointly tunes the PASS parameters and schedules the mobile devices for minimal energy consumption in OTA-FL. We study the efficiency of the proposed design and compare it against the conventional OTA-FL setting with MIMO server. Numerical experiments demonstrate that using a single-waveguide PASS at the server within a moderately sized area, the required energy for model aggregation is drastically reduced as compared to the case with fully-digital MIMO server. This introduces PASS as a potential technology for energy-efficient distributed learning in next generations of wireless systems.

cs.IT↗

arXiv2026

Regularized Top-$k$: A Bayesian Framework for Gradient Sparsification

Error accumulation is effective for gradient sparsification in distributed settings: initially-unselected gradient entries are eventually selected as their accumulated error exceeds a certain level. The accumulation essentially behaves as a scaling of the learning rate for the selected entries. Although this property prevents the slow-down of lateral movements in distributed gradient descent, it can deteriorate convergence in some settings. This work proposes a novel sparsification scheme that controls the learning rate scaling of error accumulation. The development of this scheme follows two major steps: first, gradient sparsification is formulated as an inverse probability (inference) problem, and the Bayesian optimal sparsification mask is derived as a maximum-a-posteriori estimator. Using the prior distribution inherited from Top-k, we derive a new sparsification algorithm which can be interpreted as a regularized form of Top-k. We call this algorithm regularized Top-k (RegTop-k). It utilizes past aggregated gradients to evaluate posterior statistics of the next aggregation. It then prioritizes the local accumulated gradient entries based on these posterior statistics. We validate our derivation through various numerical experiments. In distributed linear regression, it is observed that while Top-k remains at a fixed distance from the global optimum, RegTop-k converges to the global optimum at significantly higher compression ratios. We further demonstrate the generalization of this observation by employing RegTop-k in distributed training of ResNet-18 on CIFAR-10, as well as fine-tuning of multiple computer vision models on the ImageNette dataset. Our numerical results confirm that as the compression ratio increases, RegTop-k sparsification noticeably outperforms Top-k.

cs.LG↗

arXiv2025

Coding for Computation: Efficient Compression of Neural Networks for Reconfigurable Hardware

As state of the art neural networks (NNs) continue to grow in size, their resource-efficient implementation becomes ever more important. In this paper, we introduce a compression scheme that reduces the number of computations required for NN inference on reconfigurable hardware such as FPGAs. This is achieved by combining pruning via regularized training, weight sharing and linear computation coding (LCC). Contrary to common NN compression techniques, where the objective is to reduce the memory used for storing the weights of the NNs, our approach is optimized to reduce the number of additions required for inference in a hardware-friendly manner. The proposed scheme achieves competitive performance for simple multilayer perceptrons, as well as for large scale deep NNs such as ResNet-34.

cs.LG↗

arXiv2025

MIMO-PASS: Uplink and Downlink Transmission via MIMO Pinching-Antenna Systems

Pinching-antenna systems (PASSs) are a recent flexible-antenna technology that is realized by attaching simple components, referred to as pinching elements, to dielectric waveguides. This work explores the potential of deploying PASS for uplink and downlink transmission in multiuser MIMO settings. For downlink PASS-aided communication, we formulate the optimal hybrid beamforming, in which the digital precoding matrix at the access point and the location of pinching elements on the waveguides are jointly optimized to maximize the achievable weighted sum-rate. Invoking fractional programming and Gauss-Seidel approach, we propose two low-complexity algorithms to iteratively update the precoding matrix and activated locations of the pinching elements. We further study uplink transmission aided by a PASS, where an iterative scheme is designed to address the underlying hybrid multiuser detection problem. We validate the proposed schemes through extensive numerical experiments. The results demonstrate that using a PASS, the throughput in both uplink and downlink is boosted significantly as compared with baseline MIMO architectures, such as massive MIMO~and classical hybrid analog-digital designs. This highlights the great potential of PASSs, making it a promising reconfigurable antenna technology for next-generation wireless systems.

cs.IT↗

arXiv2025

Downlink Beamforming with Pinching-Antenna Assisted MIMO Systems

Pinching antennas have been recently proposed as a promising flexible-antenna technology, which can be implemented by attaching low-cost pinching elements to dielectric waveguides. This work explores the potential of employing pinching antenna systems (PASs) for downlink transmission in a multiuser MIMO setting. We consider the problem of hybrid beamforming, where the digital precoder at the access point and the activated locations of the pinching elements are jointly optimized to maximize the achievable weighted sum-rate. Invoking fractional programming, a novel low-complexity algorithm is developed to iteratively update the precoding matrix and the locations of the pinching antennas. We validate the proposed scheme through extensive numerical experiments. Our investigations demonstrate that using PAS the system throughput can be significantly boosted as compared with the conventional fixed-location antenna systems, enlightening the potential of PAS as an enabling candidate for next-generation wireless networks.

cs.IT↗

arXiv2025

Over-the-Air Fair Federated Learning via Multi-Objective Optimization

In federated learning (FL), heterogeneity among the local dataset distributions of clients can result in unsatisfactory performance for some, leading to an unfair model. To address this challenge, we propose an over-the-air fair federated learning algorithm (OTA-FFL), which leverages over-the-air computation to train fair FL models. By formulating FL as a multi-objective minimization problem, we introduce a modified Chebyshev approach to compute adaptive weighting coefficients for gradient aggregation in each communication round. To enable efficient aggregation over the multiple access channel, we derive analytical solutions for the optimal transmit scalars at the clients and the de-noising scalar at the parameter server. Extensive experiments demonstrate the superiority of OTA-FFL in achieving fairness and robust performance compared to existing methods.

cs.LG↗

arXiv2024

GP-FL: Model-Based Hessian Estimation for Second-Order Over-the-Air Federated Learning

Second-order methods are widely adopted to improve the convergence rate of learning algorithms. In federated learning (FL), these methods require the clients to share their local Hessian matrices with the parameter server (PS), which comes at a prohibitive communication cost. A classical solution to this issue is to approximate the global Hessian matrix from the first-order information. Unlike in idealized networks, this solution does not perform effectively in over-the-air FL settings, where the PS receives noisy versions of the local gradients. This paper introduces a novel second-order FL framework tailored for wireless channels. The pivotal innovation lies in the PS's capability to directly estimate the global Hessian matrix from the received noisy local gradients via a non-parametric method: the PS models the unknown Hessian matrix as a Gaussian process, and then uses the temporal relation between the gradients and Hessian along with the channel model to find a stochastic estimator for the global Hessian matrix. We refer to this method as Gaussian process-based Hessian modeling for wireless FL (GP-FL) and show that it exhibits a linear-quadratic convergence rate. Numerical experiments on various datasets demonstrate that GP-FL outperforms all classical baseline first and second order FL approaches.

cs.LG↗